The purpose of the proposed study is to measure the ability of normally hearing and hearing-impaired listeners to discriminate between sounds on the basis of the shape or contour of the power spectra of the sounds. This capability, often called "auditory profile analysis," is thought to underlie the discrimination and identification of many types of sounds - including speech sounds. There have been many past studies of sound detection or discrimination in which the signal created a difference in spectral shape, even some studies purporting to study "spectral shape discrimination" as a separate form of auditory analysis. It has only been quite recently, however, that researchers have come to appreciate the extent to which the responses of observers in such studies may reflect simple comparisons of sound intensity rather than of the spectral pattern of the sounds and, conversely, the extent to which the judgments of listeners in many traditional detection tasks may reflect decisions about spectral shape. The separation of detection strategies based on differences in spectral shape from differences in overall intensity or the intensity within a single "critical band," is possible by a procedure first described by Spiegel, Picardi and Green (J. Acoust. Soc. Am., v70, 1981) in which the overall level of the sounds is randomized. In that procedure, the statistical properties limit performance based on differences in sound level. We propose a series of experiments which will measure the just- discriminable difference in spectral shape in normally hearing listeners and in listeners with sensorineural hearing impairment. In each of these experiments, the results will be analyzed with respect to the statistical constraints of the procedure, and, in some experiments to the capability of the individual listener to discriminate between sounds based solely on sound intensity. In addition, the experiments with, normally hearing listeners are designed to evaluate certain aspects of a model of some of the factors we believe are important in spectral shape discrimination.